Over the years, liquid chromatography tandem mass spectrometry (LC-MS/MS) has become the pharmaceutical industry standard for analysis of in vivo and in vitro samples.[3, 6, 7, 8] This is due to the speed, selectivity, and sensitivity the combination gives. Even though speed, selectivity, and sensitivity has become increasingly better, even in just the past 5 years, the need for going faster, having lower limits of detection, and analyzing more compounds at one time is always there.
To accommodate the need for faster separations, ultra high-pressure liquid chromatography (UHPLC) has become increasingly popular compared to the conventional high-pressure liquid chromatography (HPLC).[3, 6] UHPLC allows for the use of smaller particle size columns, without system failure due to high pressure. The use of smaller particles allows one to scale the dimensions of the column to maintain the separation already achieved under HPLC conditions, but with a faster run time. The smaller particles can also be utilized with the same or longer size column to get better separation.
Setting the mass spectrometer (MS) to do multiple reaction monitoring (MRM), along with UHPLC, helps achieve the selectivity needed.[3, 6, 8, 22] The MRM mode analyzes the sample for a specific parent and daughter combination by allowing only those particular transitions through the quadrapoles of the MS. This lets the scientist see only the peaks in the chromatogram for the compound of interest.
UHPLC has also helped with sensitivity, allowing for taller, sharper peaks compared to peaks seen under HPLC conditions.[3] The main sensitivity gains, however, have come from better ionization in the MS source, getting more analyte into the orifice of the MS, and better efficiency in the detector changing analyte to signal. Even with these improvements in sensitivity, it never seems to be enough, especially when matrix effects are taken into consideration.[3, 6, 7]
In simplest terms, matrix effects are due to the analyte co-eluting with an endogenous matrix compenent (e.g. phospholipid in animal plasma) that either enhances or suppresses the signal of the analyte of interest.[3] This is only a concern if the samples are of a different matrix, (e.g. a different lot of the same species) than the standards or if the analyte of interest doesn't respond consistently.
Matrix interferences can be reduced with extensive sample preparation. The three main methods for sample preparation are protein precipitation (PPE), liquid-liquid (LLE), and solid-phase (SPE) extraction. While LLE and SPE are more extensive and result in cleaner sample extract than PPE, the industry standard for sample preparation is PPE due to the speed at which samples can be extracted. [3, 6]
Two reasons that matrix effects are a concern are not knowing whether the suppression or enhancement is the same from one sample to the next and the possibility that suppression will increase the lower limit of detection on the instrument. Generally speaking, phospholipids elute later than small molecules in the chromatogram due to their greasy nature.[3] Salts and some dose vehicles elute towards the early part of the chromatogram because they don't retain on column.[6] Since PPE is typically used, especially in high-throughput environments, the samples will not be clean of things like dosing vehicles, salts, and phospholipids. Since these will be left in the samples, the best course of action is to alter the retention of the analyte to the middle of the chromatogram.
Typical conditions run on UHPLC-MS/MS are gradients going from a high aqueous percentage to a high organic percentage over just a couple of minutes.[3] The mobile phase is typically modified by an acidic additive, such as formic acid.[7] Under these typical low pH conditions, the basic compounds will be protonated, and elute first, while the acidic compounds will be neutral and elute later.[3] Under these conditions, the more basic compounds need to be moved away from the salts at the beginning of the run and the acidic compounds moved away from the phospholipids eluting later in the run.
To help with this, the pH of the mobile phase can be raised. Instead of utilizing formic acid as the mobile phase modifier, a basic modifier, such as ammonia hydroxide (NH3OH) can be used. One of many examples showing how altering the pH can have a beneficial effect on matrix effects comes from looking at phenobarbital. Under acidic conditions, phenobarbital co-elutes with a variety of endogenous species of human serum.[17] When high pH is utilized, the phenobarbital is completely resolved from these serum components.[17]
Figure 1. At pH 4.4, phenobarbital (PB) is barely visible from the endogenous species of human serum. At pH 8.23, PB elutes earlier and is completely resolved.[17]
A more acidic mobile phase causes bases to elute earlier than a higher pH mobile phase.[1] At a higher pH, the basic compounds will be unprotonated when interacting with the column and elute later,[3, 22] where as the acidic compounds will be protonated and interact less with the column, causing them to elute earlier.[10, 22] In Figure 2 below, the pKa is at the inflection point of the curve, a neutral pH. The higher the pH is, the less likely one would find the analyte charged. The lower the pH, the more likely one would find the analyte in its ionized form. At a pH within 2 units of the pKa, half of the analyte will be in its ionized form and the other half in its unionized form.
Figure 2. Retention of a theoretical analyte as a function of pH.[10]
From this, one can draw the conclusion that an analyte’s retention is dependent on the pKa of that analyte and of the pH of the mobile phase, as well as other factors (e.g. stationary phase of the column). The combination of the pH and pKa will alter the interaction the analyte will have with the stationary phase on column. An analyte on a C18 column at a pH higher than the pKa of the analyte will retain longer. This is because there will be more hydrophobic interactions with the C18 stationary phase.[10] At a pH lower than the pKa of the analyte, the analyte becomes ionized and has fewer interactions with the C18 stationary phase, causing the analyte to elute earlier. [10] To show how this works, Delatour and Leclercq looked at a mixture of eleven pharmaceutical compounds.[19] High pH conditions were able to resolve the analytes much better than low pH conditions. As can bee seen by comparing the pKa value from Figure 3 with the chromatogram in Figure 4, the more basic analytes increased retention until the pH near the pKa was reached.
Figure 3. Compounds analyzed by Delatour and Leclercq.[19]
Figure 4. LC-MS chromatograms at varying pH of compounds in figure 3.[19]
One thing to keep in mind about the pKa of analytes, whether basic or acidic, is that the pKa will shift depending on the organic content of the mobile phase. As the organic component of the mobile phase increases, the pKa of basic compounds decreases and the pKa of acidic compounds increases.[10] When the mobile phase pH approaces the pKa of the analyte of interest, the peak shape tends to become irregular. To help prevent the irregular peak shape, the pH of the mobile phase should be buffered at an extreme pH (e.g. pH >10).
Krummen and Frei examined the separation of five peptides under various pH conditions.[9] They found that when the pH was increased to a value greater than 9 could they get the separation needed. At a pH greater than 9, the nitrogen of the amino group was sure to be deprotonated. Krummen and Frei also noted that above a pH of 9, the organic solvent utilized didn’t alter the retention of their analytes because the pKa of their analytes was always lower than the pH of the mobile phase.[9]
Switching to basic pH conditions over acidic comes with concerns on the LC side of the analysis. Research done by Smith and Evans showed that while utilizing a mobile phase at pH 9.8, only one of the 9 tested compounds eluted.[2] They conjectured that the reason was the interactions between the analyte and the unreacted silanols in the stationary phase. Depending on the pKa of the analytes used by Smith and Evans, the basic site on the analyte could have still been protonated and interacted with the ionized silanol groups.
There are many different factors that affect the ability of the column to perform under high pH conditions without degrading quickly. Firstly, the support for the bonded phase of the column must be considered. Using silica for the support can create issues (as mentioned above) with interactions between the ionized silanol and the analyte of interest.[12] Generally, using a higher pH would help to minimize these interactions for basic compounds since the compounds would exist in their neutral form. Other support types exist, but these have issues with reproducibility, efficiency, and types of mobile phases that can be utilized.[12] Thus, using silica as the basis of the stationary phase is the most attractive option.
After selecting the support, the next part of the column to consider is what is bonded to the support. While different phases are needed for different reasons, the silica receives increased shileding when bulkier groups are bonded. Kirkland et al. compared a C8 column to a C18 column and showed that the peaks coming from the C8 column become unsymmetrical well before those on the C18 columns.[11]
Figure 5. Peak asymmetry as a function of column volumes of purge for a C8 and C18 column. The C8 column peak asymmetry increased rapidly compared to the C18.[11]
In recent years, column manufacturers have increasingly endcapped their columns. To help protect the silica base, ligands can be attached at the ends of the column stationary phase to endcap it, protecting the most commonly unreacted silanols from interaction with the analytes.[15] Kirkland et al. performed experiments that showed that not end-capping the column caused degradation of the column to occur much faster than when the column was end-capped.[12] Their experiments suggest that the ligands create additional protection by lowering the surface energy and increasing the surface tension.
With columns being manufactured to handle the high pH mobile phases, the scientist must choose the main component to the mobile phase. Typical mobile phases are acetonitrile and methanol. Findings from Kirkland et al. in 1998 showed that methanol was better than acetonitrile at maintaining the column integrity.[11]
Studies have also shown that adding a buffer to the mobile phase helps to slow the degradation of the column. Buffers with amines seem to have the most effect on slowing column degradation. The nitrogen of the amine can bind to the ionized silica, leaving the rest of the buffer molecule to act as a shield on the surface of the silica.[11, 12]
Figure 6. Additives with a nitrogen group can attach to the colum and help in shielding. A bulkier molecule does a better job of shielding the unreacted silica portion of the column.[11]
One of the last options the scientist has to help slow column degradation is to utilize a precolumn. The intent of the precolumn is to saturate the mobile phase with degraded silica prior to entering the column. As long as the mobile phase is saturated prior to the column, the degradation of the silica on column should be drastically reduced.[11] Utilizing a precolumn without a bonded phase, as opposed to a precolumn of the same bonded phase as the column, will allow for the most saturation of the mobile phase with silica; thus having the best effect on slowing column degradation.
Taking everything above into consideration, high pH can easily be used with LC instrumentation. Classical theory, however, suggests that this is not the case for MS instrumentation. In the pharmaceutical industry, the two most utilized ionization sources are atmospheric pressure chemical ionization (APCI) and electrospray ionization (ESI);[7, 8] with ESI being the most utilized due to the enhanced sensitivity it shows over APCI.[8]
ESI works by converting the LC eluent to a mist of charged droplets. These droplets are evaporated, typically by nitrogen, until they reach a point of charge instability where the droplets then break in to smaller droplets.[7] The droplets continue to evaporate and break in this fashion until the only thing left is a charged molecule.[8] While molecules can be multiply charged, those less than 1000 amu are generally singly charged. Ammonium acetate and formic acid are common additives that help promote protonation of molecules.[7]
Many add formic acid to the mobile phase because it is widely believed that the mass spectrum generated is representative of the ions that exist prior to the electrospray process.[23] Because ions are forming prior to the source, it should be predicted that the signal intensity should be altered as the mobile phase pH is altered.[21] Basic analytes especially should decrease in intensity as the pH is increased because more of the molecules are in their unprotonated form, as discussed earlier.
Completely contradicting this school of thought, basic analytes still have signal when a high pH is utilized, and the signal, in most cases, is actually higher than when compared to the same compounds uner low pH conditoins. Looking at work done by Peng and Farkas,[14] one can see how the response changed with varying the pH. Figure 7 and Figure 8 demonstrate with 14 compounds that, under normal conditions (0.1% formic acid), the response is lower than when running at any of the various higher pH values, except in two cases were the response is similar.
Figure 7. Response of 14 compounds at various pH.[14]
Figure 8. Chromatographs of the 14 compounds in figure 7. The darker shaded peak was run at pH 10 with 10mM ammonium carbonate. The lighter shaded peak was run with 0.1% formic acid.[14]
This phenomenon does not occur with just small pharmaceutical molecules either. Tomlinson and Chicz look at peptides under basic mobile phase conditions.[4] Their results were just as astonishing. Figure 9 shows in increase of about 15 fold when running basic pH compared to acidic pH for a doubly charged peptide.
Figure 9. Doubly charged peptide. Spectra A comes from a mobile phase at pH 3.1. Spectra B comes from a mobile phase with at pH 10.0 with ammonium hydroxide as the additive.[4]
One problem that typically occurs when working with peptides is high background noise, as was seen in Figure 9. Changing to a basic pH environment elimated a lot of the background noise allowing Tomlinson and Chicz to detect many more peptide peaks as shown in Figure 10. This lower background allowed them to identify over 100 peptides compared to only 5 under acidic conditions.[4]
Figure 10. Chromatogram A came from a mobile phase at pH 3.1 containing two additives: ammonium acetate and acetic acid. Chromatogram B came from a mobile phase at pH 10 containing the additive ammonium hydroxide. The mobile phase at pH 10 allowed for a larger elution window of the peptide, allowing for more peaks to be analyzed.[4]
While Tomlinson and Chicz used acetic acid as their mobile phase additive, that is not typical when working with peptides. Trifluoroacetic acid (TFA) is one of the more common additives because it helps provide excellent peak shape.[5] The main issue with TFA is that it drastically reduces the signal in the MS. Experiments were performed by Stokvis et al. to show that utilizing a mixture of 10mM aqueous ammonia in acetonitrile not only provided peak shape similar to that of TFA, but an intense signal gain over 35 fold.[5]
Figure 11. All signal gain was compared to a mobile phase containing 0.04% TFA. The post-column mixing was done with propionic acid and isopropanol in a ration of 75:25 v:v.[5]
Berg et al. looked at opiates and cocaine in urine and witnessed a 3-4 fold increase in most cases when comparing peak area in high pH to low pH conditions.[13]
Running the mobile phase with acidic modifiers can create ionization inefficiencies in negative mode, causing signal suppression. Schaefer and Dixon Jr. looked at this phenomenon under APCI conditions.[16] As they explain, APCI works differently than ESI. APCI causes anything in the gas phase to undergo reactions started by a corona discharge. Ionization under APCI typically happens through proton transfer reactions. The species in the gas phase that has the higher proton affinity will gain the proton. Because of how APCI works, the pH of the mobile phase should have no effect on the ionization process. Therefore, the same increases gained running in positive ion spray with high pH can also be realized with negative ion spray and high pH.
The chemical nature of the analytes does, however, affect the ionization process. Schaefer and Dixon Jr. tested this.[16] They found that acidic additives, especially formic acid, reduce the formation of negative analyte ions. When a base, such as N-methylmorpholine, is utilized, the formation of negative analyte ions is increased.
Figure 12. Chromatograms at m/z 153 for 2-nonbornaneacetic acid. Mobile phases were 1:1 mixtures of acetonitile (A) 10mM N-methylmorpholine, (B) 10mM ammonium acetate, (C) 100mM ammonium acetate, and (D) 10mM formic acid. The number in the upper right of each chromatogram represents the height of the tallest peak in that chromatogram.[16]
In most every case shown, the high pH mobile phase has increased sensitivity. To determine if hihg pH mobile phase is purely a MS affect or if the LC also plays a role, basic mobile phase must be added to the rest of the LC mobile phase post column. Doing this allows for everything from the LC to remain consistent between acidic and basic conditions, isolating the sensitivity change to the post column additivies. Mess et al. tested 25 compounds.[17] Of these 25, all showed signal enhancement with the post column addition of ammonium hydroxide when compared to formic and acetic acid.
Figure 13. Signal enhancement comparison for 25 compounds of 50mM ammonium hydroxide vs 1% formic acid (dark blue bar) and 1% acetic acid (light blue bar).[18]
The test from Mess et al. undoubtedly showed that the signal enhancement is from the MS. This is surprising as, according to previous theories on how ESI works, there should be a decrease in signal as the pH rises. There have been a few attempts to explain this, but nothing has been proven.
The first attempt to describe what was happening under high pH conditions in ESI was from Mansoori et al. who coined the term “wrong-way-round” ionization.[20] The “right-way-round” is the formation of protonated ions in the mobile phase by the acid, whereas the “wrong-way-round” is the formation of protonated ions in the gas phase. To look at this, Mansoori et al. looked at precursor ions being formed.[20] Under acidic conditions, adducts of the protonated ion were formed with water, methanol, and acetic acid. Under basic conditions, adducts of the analyte with an ammonium ion and a dimer ion were seen. Mansoori et al. were interested in determing how the analyte plus ammonium and dimer ions become protonated analyte ions.
The most common theory incorporates both ion-molecule reactions in the gas phase as well as collision induced dissociation (equations 1 and 2).[21]
(1)
(2)
Others have proposed that proton transfer actually takes place in the liquid phase of the droplet with the evaporation of ammonia.[14]
(3)
Equations 1, 2, and 3 work well to explain the process when ammonium is present in the mobile phase additive. The equations break down when ammonium hydroxide is replaced with a modifier like sodium hydroxide. A significant signal is still seen with sodium hydroxide, but there is no ammonium present for ESI to work by the previous equations. Zhou and Cook recognize the following pathway, similar to that of APCI:[21]
(4)
(5)
There is yet another common thought for the formation of the protonated ion in basic mobile phase. This time, the authors try to utilize a process that can occur with both ammonia containing and non-ammonia containing additives:[19, 20]
(6)
It could be some, none, or any combination of the equations above that explain why a signal is seen when a basic pH mobile phase is utilized. Whatever the case, a signal is seen and this signal is generally larger than the signal running under the typical acidic conditions.
In reviewing all of these works, one conclusion can be drawn: high pH mobile phase works just as well if not better than the typical low pH mobile phase conditions. With new packing technology, columns are gaining a wider pH range under which they can work, as opposed to the columns from the late 1990s that quickly degraded over a pH of 7.[16] The hang-up to switching over to high pH over the typical low pH conditions seems to be coming from the conventional stand point that acidified mobile phases will help in the formation of protonated ions. It will take time to overcome these conventional thoughts and have the analytical community switch gears, as pointed out by these and many other papers, and run under the generally more sensitive basic conditions.
References
[1] Kaliszan R., Wiczling P., and Markuszewski M. pH gradient high-performance liquid chromatography: theory and applications. Journal of Chromatography A 2004, 1060(1-2): 165–175. doi: http://dx.doi.org/10.1016/j.chroma.2004.04.081
[2] Smith N. and Evans M. The efficient analysis of neutral and highly polar pharmaceutical compounds using reversed-phase and ion-exchange electrochromatography. Chromatographia 1995, 41(5-6): 197-203. doi: http://dx.doi.org/10.1007/BF02688025
[3] Chambers E., Wagrowski-Diehl D., Lu Z., and Mazzeo J. Systematic and comprehensive strategy for reducing matrix effects in LC/MS/MS analyses. Journal of Chromatography B 2007, 852(1-2): 22–34. doi: http://dx.doi.org/10.1016/j.jchromb.2006.12.030
[4] Tomlinson A. and Chicz R. Microcapillary liquid chromatography/tandem mass spectrometry using alkaline pH mobile phases and positive ion detection. Rapid Commun. Mass Spectrom. 2003, 17(9): 909–916. doi: http://dx.doi.org/10.1002/rcm.1001
[5] Stokvis E., Rosing H., López-Lázaro L., Rodriguez I., Jimeno J., Supko J., Schellens J., and Beijnen J. Quantitative analysis of the novel depsipeptide anticancer drug Kahalalide F in human plasma by high-performance liquid chromatography under basic conditions coupled to electrospray ionization tandem mass spectrometry. J. Mass Spectrom. 2002, 37(9): 992–1000. doi: http://dx.doi.org/10.1002/jms.362
[6] Xu R., Fan L., Rieser M., and El-Shourbagy T. Recent advances in high-throughput quantitative bioanalysis by LC–MS/MS. Journal of Pharmaceutical and Biomedical Analysis 2007, 44(2): 342–355. doi: http://dx.doi.org/10.1016/j.jpba.2007.02.006
[7] Lim C. and Lord G. Current Developments in LC-MS for Pharmaceutical Analysis. Biological and Pharmaceutical Bulletin 2002, 25(5): 547-557. doi: http://dx.doi.org/10.1248/bpb.25.547
[8] Ermer J. and Vogel M. Applications of hyphenated LC-MS techniques in pharmaceutical analysis. Biomed. Chromatogr. 2000, 14(6): 373–383. doi: http://dx.doi.org/10.1002/1099-0801(200010)14:6<373::AID-BMC29>3.0.CO;2-S
[9] Krummen K. and Frei R. The separation of nonapeptides by reversed-phase high-performance liquid chromatography. Journal of Chromatography A 1977, 132(1): 27–36. doi: http://dx.doi.org/10.1016/S0021-9673(00)93767-1
[10] LoBrutto R., Jones A., Kazakevich Y., and McNair H. Effect of the eluent pH and acidic modifiers in high-performance liquid chromatography retention of basic analytes. Journal of Chromatography A 2001, 913(1-2): 173–187. doi: http://dx.doi.org/10.1016/S0021-9673(00)01012-8
[11] Kirkland J., van Straten M., and Claessens H. Reversed-phase high-performance liquid chromatography of basic compounds at pH 11 with silica-based column packings. Journal of Chromatography A 1998, 797(1-2): 111–120. doi: http://dx.doi.org/10.1016/S0021-9673(97)00865-0
[12] Kirkland J., Henderson J., DeStefano J., van Straten M., and Claessens H. Stability of silica-based, endcapped columns with pH 7 and 11 mobile phases for reversed-phase high-performance liquid chromatography. Journal of Chromatography A 1997, 762(1-2): 97–112. doi: http://dx.doi.org/10.1016/S0021-9673(96)00945-4
[13] Berg T., Lundanes E., Christophersen A., and Strand D. Determination of opiates and cocaine in urine by high pH mobile phase reversed phase UPLC–MS/MS. Journal of Chromatography B 2009, 877(4): 421–432. doi: http://dx.doi.org/10.1016/j.jchromb.2008.12.052
[14] Peng L. and Farkas T. Analysis of basic compounds by reversed-phase liquid chromatography–electrospray mass spectrometry in high-pH mobile phases. Journal of Chromatography A 2008, 1179(2): 131–144. doi: http://dx.doi.org/10.1016/j.chroma.2007.11.048
[15] Sýkora D., Tesařová E., and Popl M. Interactions of basic compounds in reversed-phase high-performance liquid chromatography influence of sorbent character, mobile phase composition, and pH on retention of basic compounds. Journal of Chromatography A 1997, 758(1): 37–51. doi: http://dx.doi.org/10.1016/S0021-9673(96)00691-7
[16] Schaefer W. and Dixon Jr. F. Effect of high-performance liquid chromatography mobile phase components on sensitivity in negative atmospheric pressure chemical ionization liquid chromatography-mass spectrometry. Journal of the American Society for Mass Spectrometry 1996, 7(10): 1059-1069. doi: http://dx.doi.org/10.1016/1044-0305(96)00049-9
[17] Lee D. Reversed-Phase HPLC from pH 1 to 13. J. Chromatogr. Sci. 1982, 20(5): 203-208. Link
[18] Mess J., Lahaie M., Furtado M., and Garofolo F. Effect of high pH mobile phase on the sensitivity of multiple drugs by LC positive electrospray ionization MS/MS. Bioanalysis 2009, 1(8):1419-30. Link
[19] Delatour C. and Leclercq L. Positive electrospray liquid chromatography/mass spectrometry using high-pH gradients: a way to combine selectivity and sensitivity for a large variety of drugs. Rapid Commun. Mass Spectrom. 2005, 19(10): 1359–1362. doi: http://dx.doi.org/10.1002/rcm.1923
[20] Mansoori B., Volmer D., and Boyd R. ‘Wrong-way-round’ Electrospray Ionization of Amino Acids. Rapid Commun. Mass Spectrom. 1997, 11(10): 1120–1130. doi: http://dx.doi.org/10.1002/(SICI)1097-0231(19970630)11:10<1120::AID-RCM976>3.0.CO;2-Q
[21] Zhou S. and Cook K. Protonation in electrospray mass spectrometry: Wrong-way-round or right-way-round? Journal of the American Society for Mass Spectrometry 2000, 11(11): 961-966 doi: http://dx.doi.org/10.1016/S1044-0305(00)00174-4
[22] Cheng Y., Lu Z., and Neue, U. Ultrafast liquid chromatography/ultraviolet and liquid chromatography/tandem mass spectrometric analysis. Rapid Commun. Mass Spectrom. 2001, 15(2): 141–151. doi: http://dx.doi.org/10.1002/1097-0231(20010130)15:2<141::AID-RCM201>3.0.CO;2-I
[23] Wang G. and Cole R. Effect of Solution Ionic Strength on Analyte Charge State Distributions in Positive and Negative Ion Electrospray Mass Spectrometry. Anal. Chem 1994, 66 (21): 3702–3708. doi: http://dx.doi.org/10.1021/ac00093a026
James Marr
Over the years, liquid chromatography tandem mass spectrometry (LC-MS/MS) has become the pharmaceutical industry standard for analysis of in vivo and in vitro samples.[3, 6, 7, 8] This is due to the speed, selectivity, and sensitivity the combination gives. Even though speed, selectivity, and sensitivity has become increasingly better, even in just the past 5 years, the need for going faster, having lower limits of detection, and analyzing more compounds at one time is always there.
To accommodate the need for faster separations, ultra high-pressure liquid chromatography (UHPLC) has become increasingly popular compared to the conventional high-pressure liquid chromatography (HPLC).[3, 6] UHPLC allows for the use of smaller particle size columns, without system failure due to high pressure. The use of smaller particles allows one to scale the dimensions of the column to maintain the separation already achieved under HPLC conditions, but with a faster run time. The smaller particles can also be utilized with the same or longer size column to get better separation.
Setting the mass spectrometer (MS) to do multiple reaction monitoring (MRM), along with UHPLC, helps achieve the selectivity needed.[3, 6, 8, 22] The MRM mode analyzes the sample for a specific parent and daughter combination by allowing only those particular transitions through the quadrapoles of the MS. This lets the scientist see only the peaks in the chromatogram for the compound of interest.
UHPLC has also helped with sensitivity, allowing for taller, sharper peaks compared to peaks seen under HPLC conditions.[3] The main sensitivity gains, however, have come from better ionization in the MS source, getting more analyte into the orifice of the MS, and better efficiency in the detector changing analyte to signal. Even with these improvements in sensitivity, it never seems to be enough, especially when matrix effects are taken into consideration.[3, 6, 7]
In simplest terms, matrix effects are due to the analyte co-eluting with an endogenous matrix compenent (e.g. phospholipid in animal plasma) that either enhances or suppresses the signal of the analyte of interest.[3] This is only a concern if the samples are of a different matrix, (e.g. a different lot of the same species) than the standards or if the analyte of interest doesn't respond consistently.
Matrix interferences can be reduced with extensive sample preparation. The three main methods for sample preparation are protein precipitation (PPE), liquid-liquid (LLE), and solid-phase (SPE) extraction. While LLE and SPE are more extensive and result in cleaner sample extract than PPE, the industry standard for sample preparation is PPE due to the speed at which samples can be extracted. [3, 6]
Two reasons that matrix effects are a concern are not knowing whether the suppression or enhancement is the same from one sample to the next and the possibility that suppression will increase the lower limit of detection on the instrument. Generally speaking, phospholipids elute later than small molecules in the chromatogram due to their greasy nature.[3] Salts and some dose vehicles elute towards the early part of the chromatogram because they don't retain on column.[6] Since PPE is typically used, especially in high-throughput environments, the samples will not be clean of things like dosing vehicles, salts, and phospholipids. Since these will be left in the samples, the best course of action is to alter the retention of the analyte to the middle of the chromatogram.
Typical conditions run on UHPLC-MS/MS are gradients going from a high aqueous percentage to a high organic percentage over just a couple of minutes.[3] The mobile phase is typically modified by an acidic additive, such as formic acid.[7] Under these typical low pH conditions, the basic compounds will be protonated, and elute first, while the acidic compounds will be neutral and elute later.[3] Under these conditions, the more basic compounds need to be moved away from the salts at the beginning of the run and the acidic compounds moved away from the phospholipids eluting later in the run.
To help with this, the pH of the mobile phase can be raised. Instead of utilizing formic acid as the mobile phase modifier, a basic modifier, such as ammonia hydroxide (NH3OH) can be used. One of many examples showing how altering the pH can have a beneficial effect on matrix effects comes from looking at phenobarbital. Under acidic conditions, phenobarbital co-elutes with a variety of endogenous species of human serum.[17] When high pH is utilized, the phenobarbital is completely resolved from these serum components.[17]
Figure 1. At pH 4.4, phenobarbital (PB) is barely visible from the endogenous species of human serum. At pH 8.23, PB elutes earlier and is completely resolved.[17]
A more acidic mobile phase causes bases to elute earlier than a higher pH mobile phase.[1] At a higher pH, the basic compounds will be unprotonated when interacting with the column and elute later,[3, 22] where as the acidic compounds will be protonated and interact less with the column, causing them to elute earlier.[10, 22] In Figure 2 below, the pKa is at the inflection point of the curve, a neutral pH. The higher the pH is, the less likely one would find the analyte charged. The lower the pH, the more likely one would find the analyte in its ionized form. At a pH within 2 units of the pKa, half of the analyte will be in its ionized form and the other half in its unionized form.
Figure 2. Retention of a theoretical analyte as a function of pH.[10]
From this, one can draw the conclusion that an analyte’s retention is dependent on the pKa of that analyte and of the pH of the mobile phase, as well as other factors (e.g. stationary phase of the column). The combination of the pH and pKa will alter the interaction the analyte will have with the stationary phase on column. An analyte on a C18 column at a pH higher than the pKa of the analyte will retain longer. This is because there will be more hydrophobic interactions with the C18 stationary phase.[10] At a pH lower than the pKa of the analyte, the analyte becomes ionized and has fewer interactions with the C18 stationary phase, causing the analyte to elute earlier. [10] To show how this works, Delatour and Leclercq looked at a mixture of eleven pharmaceutical compounds.[19] High pH conditions were able to resolve the analytes much better than low pH conditions. As can bee seen by comparing the pKa value from Figure 3 with the chromatogram in Figure 4, the more basic analytes increased retention until the pH near the pKa was reached.
Figure 3. Compounds analyzed by Delatour and Leclercq.[19]
Figure 4. LC-MS chromatograms at varying pH of compounds in figure 3.[19]
One thing to keep in mind about the pKa of analytes, whether basic or acidic, is that the pKa will shift depending on the organic content of the mobile phase. As the organic component of the mobile phase increases, the pKa of basic compounds decreases and the pKa of acidic compounds increases.[10] When the mobile phase pH approaces the pKa of the analyte of interest, the peak shape tends to become irregular. To help prevent the irregular peak shape, the pH of the mobile phase should be buffered at an extreme pH (e.g. pH >10).
Krummen and Frei examined the separation of five peptides under various pH conditions.[9] They found that when the pH was increased to a value greater than 9 could they get the separation needed. At a pH greater than 9, the nitrogen of the amino group was sure to be deprotonated. Krummen and Frei also noted that above a pH of 9, the organic solvent utilized didn’t alter the retention of their analytes because the pKa of their analytes was always lower than the pH of the mobile phase.[9]
Switching to basic pH conditions over acidic comes with concerns on the LC side of the analysis. Research done by Smith and Evans showed that while utilizing a mobile phase at pH 9.8, only one of the 9 tested compounds eluted.[2] They conjectured that the reason was the interactions between the analyte and the unreacted silanols in the stationary phase. Depending on the pKa of the analytes used by Smith and Evans, the basic site on the analyte could have still been protonated and interacted with the ionized silanol groups.
There are many different factors that affect the ability of the column to perform under high pH conditions without degrading quickly. Firstly, the support for the bonded phase of the column must be considered. Using silica for the support can create issues (as mentioned above) with interactions between the ionized silanol and the analyte of interest.[12] Generally, using a higher pH would help to minimize these interactions for basic compounds since the compounds would exist in their neutral form. Other support types exist, but these have issues with reproducibility, efficiency, and types of mobile phases that can be utilized.[12] Thus, using silica as the basis of the stationary phase is the most attractive option.
After selecting the support, the next part of the column to consider is what is bonded to the support. While different phases are needed for different reasons, the silica receives increased shileding when bulkier groups are bonded. Kirkland et al. compared a C8 column to a C18 column and showed that the peaks coming from the C8 column become unsymmetrical well before those on the C18 columns.[11]
Figure 5. Peak asymmetry as a function of column volumes of purge for a C8 and C18 column. The C8 column peak asymmetry increased rapidly compared to the C18.[11]
In recent years, column manufacturers have increasingly endcapped their columns. To help protect the silica base, ligands can be attached at the ends of the column stationary phase to endcap it, protecting the most commonly unreacted silanols from interaction with the analytes.[15] Kirkland et al. performed experiments that showed that not end-capping the column caused degradation of the column to occur much faster than when the column was end-capped.[12] Their experiments suggest that the ligands create additional protection by lowering the surface energy and increasing the surface tension.
With columns being manufactured to handle the high pH mobile phases, the scientist must choose the main component to the mobile phase. Typical mobile phases are acetonitrile and methanol. Findings from Kirkland et al. in 1998 showed that methanol was better than acetonitrile at maintaining the column integrity.[11]
Studies have also shown that adding a buffer to the mobile phase helps to slow the degradation of the column. Buffers with amines seem to have the most effect on slowing column degradation. The nitrogen of the amine can bind to the ionized silica, leaving the rest of the buffer molecule to act as a shield on the surface of the silica.[11, 12]
Figure 6. Additives with a nitrogen group can attach to the colum and help in shielding. A bulkier molecule does a better job of shielding the unreacted silica portion of the column.[11]
One of the last options the scientist has to help slow column degradation is to utilize a precolumn. The intent of the precolumn is to saturate the mobile phase with degraded silica prior to entering the column. As long as the mobile phase is saturated prior to the column, the degradation of the silica on column should be drastically reduced.[11] Utilizing a precolumn without a bonded phase, as opposed to a precolumn of the same bonded phase as the column, will allow for the most saturation of the mobile phase with silica; thus having the best effect on slowing column degradation.
Taking everything above into consideration, high pH can easily be used with LC instrumentation. Classical theory, however, suggests that this is not the case for MS instrumentation. In the pharmaceutical industry, the two most utilized ionization sources are atmospheric pressure chemical ionization (APCI) and electrospray ionization (ESI);[7, 8] with ESI being the most utilized due to the enhanced sensitivity it shows over APCI.[8]
ESI works by converting the LC eluent to a mist of charged droplets. These droplets are evaporated, typically by nitrogen, until they reach a point of charge instability where the droplets then break in to smaller droplets.[7] The droplets continue to evaporate and break in this fashion until the only thing left is a charged molecule.[8] While molecules can be multiply charged, those less than 1000 amu are generally singly charged. Ammonium acetate and formic acid are common additives that help promote protonation of molecules.[7]
Many add formic acid to the mobile phase because it is widely believed that the mass spectrum generated is representative of the ions that exist prior to the electrospray process.[23] Because ions are forming prior to the source, it should be predicted that the signal intensity should be altered as the mobile phase pH is altered.[21] Basic analytes especially should decrease in intensity as the pH is increased because more of the molecules are in their unprotonated form, as discussed earlier.
Completely contradicting this school of thought, basic analytes still have signal when a high pH is utilized, and the signal, in most cases, is actually higher than when compared to the same compounds uner low pH conditoins. Looking at work done by Peng and Farkas,[14] one can see how the response changed with varying the pH. Figure 7 and Figure 8 demonstrate with 14 compounds that, under normal conditions (0.1% formic acid), the response is lower than when running at any of the various higher pH values, except in two cases were the response is similar.
Figure 7. Response of 14 compounds at various pH.[14]
Figure 8. Chromatographs of the 14 compounds in figure 7. The darker shaded peak was run at pH 10 with 10mM ammonium carbonate. The lighter shaded peak was run with 0.1% formic acid.[14]
This phenomenon does not occur with just small pharmaceutical molecules either. Tomlinson and Chicz look at peptides under basic mobile phase conditions.[4] Their results were just as astonishing. Figure 9 shows in increase of about 15 fold when running basic pH compared to acidic pH for a doubly charged peptide.
Figure 9. Doubly charged peptide. Spectra A comes from a mobile phase at pH 3.1. Spectra B comes from a mobile phase with at pH 10.0 with ammonium hydroxide as the additive.[4]
One problem that typically occurs when working with peptides is high background noise, as was seen in Figure 9. Changing to a basic pH environment elimated a lot of the background noise allowing Tomlinson and Chicz to detect many more peptide peaks as shown in Figure 10. This lower background allowed them to identify over 100 peptides compared to only 5 under acidic conditions.[4]
Figure 10. Chromatogram A came from a mobile phase at pH 3.1 containing two additives: ammonium acetate and acetic acid. Chromatogram B came from a mobile phase at pH 10 containing the additive ammonium hydroxide. The mobile phase at pH 10 allowed for a larger elution window of the peptide, allowing for more peaks to be analyzed.[4]
While Tomlinson and Chicz used acetic acid as their mobile phase additive, that is not typical when working with peptides. Trifluoroacetic acid (TFA) is one of the more common additives because it helps provide excellent peak shape.[5] The main issue with TFA is that it drastically reduces the signal in the MS. Experiments were performed by Stokvis et al. to show that utilizing a mixture of 10mM aqueous ammonia in acetonitrile not only provided peak shape similar to that of TFA, but an intense signal gain over 35 fold.[5]
Figure 11. All signal gain was compared to a mobile phase containing 0.04% TFA. The post-column mixing was done with propionic acid and isopropanol in a ration of 75:25 v:v.[5]
Berg et al. looked at opiates and cocaine in urine and witnessed a 3-4 fold increase in most cases when comparing peak area in high pH to low pH conditions.[13]
Running the mobile phase with acidic modifiers can create ionization inefficiencies in negative mode, causing signal suppression. Schaefer and Dixon Jr. looked at this phenomenon under APCI conditions.[16] As they explain, APCI works differently than ESI. APCI causes anything in the gas phase to undergo reactions started by a corona discharge. Ionization under APCI typically happens through proton transfer reactions. The species in the gas phase that has the higher proton affinity will gain the proton. Because of how APCI works, the pH of the mobile phase should have no effect on the ionization process. Therefore, the same increases gained running in positive ion spray with high pH can also be realized with negative ion spray and high pH.
The chemical nature of the analytes does, however, affect the ionization process. Schaefer and Dixon Jr. tested this.[16] They found that acidic additives, especially formic acid, reduce the formation of negative analyte ions. When a base, such as N-methylmorpholine, is utilized, the formation of negative analyte ions is increased.
Figure 12. Chromatograms at m/z 153 for 2-nonbornaneacetic acid. Mobile phases were 1:1 mixtures of acetonitile (A) 10mM N-methylmorpholine, (B) 10mM ammonium acetate, (C) 100mM ammonium acetate, and (D) 10mM formic acid. The number in the upper right of each chromatogram represents the height of the tallest peak in that chromatogram.[16]
In most every case shown, the high pH mobile phase has increased sensitivity. To determine if hihg pH mobile phase is purely a MS affect or if the LC also plays a role, basic mobile phase must be added to the rest of the LC mobile phase post column. Doing this allows for everything from the LC to remain consistent between acidic and basic conditions, isolating the sensitivity change to the post column additivies. Mess et al. tested 25 compounds.[17] Of these 25, all showed signal enhancement with the post column addition of ammonium hydroxide when compared to formic and acetic acid.
Figure 13. Signal enhancement comparison for 25 compounds of 50mM ammonium hydroxide vs 1% formic acid (dark blue bar) and 1% acetic acid (light blue bar).[18]
The test from Mess et al. undoubtedly showed that the signal enhancement is from the MS. This is surprising as, according to previous theories on how ESI works, there should be a decrease in signal as the pH rises. There have been a few attempts to explain this, but nothing has been proven.
The first attempt to describe what was happening under high pH conditions in ESI was from Mansoori et al. who coined the term “wrong-way-round” ionization.[20] The “right-way-round” is the formation of protonated ions in the mobile phase by the acid, whereas the “wrong-way-round” is the formation of protonated ions in the gas phase. To look at this, Mansoori et al. looked at precursor ions being formed.[20] Under acidic conditions, adducts of the protonated ion were formed with water, methanol, and acetic acid. Under basic conditions, adducts of the analyte with an ammonium ion and a dimer ion were seen. Mansoori et al. were interested in determing how the analyte plus ammonium and dimer ions become protonated analyte ions.
The most common theory incorporates both ion-molecule reactions in the gas phase as well as collision induced dissociation (equations 1 and 2).[21]
Others have proposed that proton transfer actually takes place in the liquid phase of the droplet with the evaporation of ammonia.[14]
Equations 1, 2, and 3 work well to explain the process when ammonium is present in the mobile phase additive. The equations break down when ammonium hydroxide is replaced with a modifier like sodium hydroxide. A significant signal is still seen with sodium hydroxide, but there is no ammonium present for ESI to work by the previous equations. Zhou and Cook recognize the following pathway, similar to that of APCI:[21]
There is yet another common thought for the formation of the protonated ion in basic mobile phase. This time, the authors try to utilize a process that can occur with both ammonia containing and non-ammonia containing additives:[19, 20]
It could be some, none, or any combination of the equations above that explain why a signal is seen when a basic pH mobile phase is utilized. Whatever the case, a signal is seen and this signal is generally larger than the signal running under the typical acidic conditions.
In reviewing all of these works, one conclusion can be drawn: high pH mobile phase works just as well if not better than the typical low pH mobile phase conditions. With new packing technology, columns are gaining a wider pH range under which they can work, as opposed to the columns from the late 1990s that quickly degraded over a pH of 7.[16] The hang-up to switching over to high pH over the typical low pH conditions seems to be coming from the conventional stand point that acidified mobile phases will help in the formation of protonated ions. It will take time to overcome these conventional thoughts and have the analytical community switch gears, as pointed out by these and many other papers, and run under the generally more sensitive basic conditions.
References
[1] Kaliszan R., Wiczling P., and Markuszewski M. pH gradient high-performance liquid chromatography: theory and applications. Journal of Chromatography A 2004, 1060(1-2): 165–175. doi: http://dx.doi.org/10.1016/j.chroma.2004.04.081
[2] Smith N. and Evans M. The efficient analysis of neutral and highly polar pharmaceutical compounds using reversed-phase and ion-exchange electrochromatography. Chromatographia 1995, 41(5-6): 197-203. doi: http://dx.doi.org/10.1007/BF02688025
[3] Chambers E., Wagrowski-Diehl D., Lu Z., and Mazzeo J. Systematic and comprehensive strategy for reducing matrix effects in LC/MS/MS analyses. Journal of Chromatography B 2007, 852(1-2): 22–34. doi: http://dx.doi.org/10.1016/j.jchromb.2006.12.030
[4] Tomlinson A. and Chicz R. Microcapillary liquid chromatography/tandem mass spectrometry using alkaline pH mobile phases and positive ion detection. Rapid Commun. Mass Spectrom. 2003, 17(9): 909–916. doi: http://dx.doi.org/10.1002/rcm.1001
[5] Stokvis E., Rosing H., López-Lázaro L., Rodriguez I., Jimeno J., Supko J., Schellens J., and Beijnen J. Quantitative analysis of the novel depsipeptide anticancer drug Kahalalide F in human plasma by high-performance liquid chromatography under basic conditions coupled to electrospray ionization tandem mass spectrometry. J. Mass Spectrom. 2002, 37(9): 992–1000. doi: http://dx.doi.org/10.1002/jms.362
[6] Xu R., Fan L., Rieser M., and El-Shourbagy T. Recent advances in high-throughput quantitative bioanalysis by LC–MS/MS. Journal of Pharmaceutical and Biomedical Analysis 2007, 44(2): 342–355. doi: http://dx.doi.org/10.1016/j.jpba.2007.02.006
[7] Lim C. and Lord G. Current Developments in LC-MS for Pharmaceutical Analysis. Biological and Pharmaceutical Bulletin 2002, 25(5): 547-557. doi: http://dx.doi.org/10.1248/bpb.25.547
[8] Ermer J. and Vogel M. Applications of hyphenated LC-MS techniques in pharmaceutical analysis. Biomed. Chromatogr. 2000, 14(6): 373–383. doi: http://dx.doi.org/10.1002/1099-0801(200010)14:6<373::AID-BMC29>3.0.CO;2-S
[9] Krummen K. and Frei R. The separation of nonapeptides by reversed-phase high-performance liquid chromatography. Journal of Chromatography A 1977, 132(1): 27–36. doi: http://dx.doi.org/10.1016/S0021-9673(00)93767-1
[10] LoBrutto R., Jones A., Kazakevich Y., and McNair H. Effect of the eluent pH and acidic modifiers in high-performance liquid chromatography retention of basic analytes. Journal of Chromatography A 2001, 913(1-2): 173–187. doi: http://dx.doi.org/10.1016/S0021-9673(00)01012-8
[11] Kirkland J., van Straten M., and Claessens H. Reversed-phase high-performance liquid chromatography of basic compounds at pH 11 with silica-based column packings. Journal of Chromatography A 1998, 797(1-2): 111–120. doi: http://dx.doi.org/10.1016/S0021-9673(97)00865-0
[12] Kirkland J., Henderson J., DeStefano J., van Straten M., and Claessens H. Stability of silica-based, endcapped columns with pH 7 and 11 mobile phases for reversed-phase high-performance liquid chromatography. Journal of Chromatography A 1997, 762(1-2): 97–112. doi: http://dx.doi.org/10.1016/S0021-9673(96)00945-4
[13] Berg T., Lundanes E., Christophersen A., and Strand D. Determination of opiates and cocaine in urine by high pH mobile phase reversed phase UPLC–MS/MS. Journal of Chromatography B 2009, 877(4): 421–432. doi: http://dx.doi.org/10.1016/j.jchromb.2008.12.052
[14] Peng L. and Farkas T. Analysis of basic compounds by reversed-phase liquid chromatography–electrospray mass spectrometry in high-pH mobile phases. Journal of Chromatography A 2008, 1179(2): 131–144. doi: http://dx.doi.org/10.1016/j.chroma.2007.11.048
[15] Sýkora D., Tesařová E., and Popl M. Interactions of basic compounds in reversed-phase high-performance liquid chromatography influence of sorbent character, mobile phase composition, and pH on retention of basic compounds. Journal of Chromatography A 1997, 758(1): 37–51. doi: http://dx.doi.org/10.1016/S0021-9673(96)00691-7
[16] Schaefer W. and Dixon Jr. F. Effect of high-performance liquid chromatography mobile phase components on sensitivity in negative atmospheric pressure chemical ionization liquid chromatography-mass spectrometry. Journal of the American Society for Mass Spectrometry 1996, 7(10): 1059-1069. doi: http://dx.doi.org/10.1016/1044-0305(96)00049-9
[17] Lee D. Reversed-Phase HPLC from pH 1 to 13. J. Chromatogr. Sci. 1982, 20(5): 203-208. Link
[18] Mess J., Lahaie M., Furtado M., and Garofolo F. Effect of high pH mobile phase on the sensitivity of multiple drugs by LC positive electrospray ionization MS/MS. Bioanalysis 2009, 1(8):1419-30. Link
[19] Delatour C. and Leclercq L. Positive electrospray liquid chromatography/mass spectrometry using high-pH gradients: a way to combine selectivity and sensitivity for a large variety of drugs. Rapid Commun. Mass Spectrom. 2005, 19(10): 1359–1362. doi: http://dx.doi.org/10.1002/rcm.1923
[20] Mansoori B., Volmer D., and Boyd R. ‘Wrong-way-round’ Electrospray Ionization of Amino Acids. Rapid Commun. Mass Spectrom. 1997, 11(10): 1120–1130. doi: http://dx.doi.org/10.1002/(SICI)1097-0231(19970630)11:10<1120::AID-RCM976>3.0.CO;2-Q
[21] Zhou S. and Cook K. Protonation in electrospray mass spectrometry: Wrong-way-round or right-way-round? Journal of the American Society for Mass Spectrometry 2000, 11(11): 961-966 doi: http://dx.doi.org/10.1016/S1044-0305(00)00174-4
[22] Cheng Y., Lu Z., and Neue, U. Ultrafast liquid chromatography/ultraviolet and liquid chromatography/tandem mass spectrometric analysis. Rapid Commun. Mass Spectrom. 2001, 15(2): 141–151. doi: http://dx.doi.org/10.1002/1097-0231(20010130)15:2<141::AID-RCM201>3.0.CO;2-I
[23] Wang G. and Cole R. Effect of Solution Ionic Strength on Analyte Charge State Distributions in Positive and Negative Ion Electrospray Mass Spectrometry. Anal. Chem 1994, 66 (21): 3702–3708. doi: http://dx.doi.org/10.1021/ac00093a026